isolate_proto_v0.5.4

• feat(isolate_proto): introduce request_timeout (#309)

fal_v1.4.1

• feat(cdn): use upload endpoint (#306)
• feat(fal-file): convert the app-ref to a path relative to the project root (#307)

fal_v1.4.0

• feat: add retry logic for file upload & fallback repository (#277)
• feat(cdn): auth with tokens and talk directly to cdn (#300)
• feat(cdn): use X-Fal-File-Name with fal_v2 (#302)
• feat: use new machine_types (#289)
• fix: supported keys (#303)
• fix: make backwards compatible with old api
• fix: correctly set num_gpus in machine requirements for fal.App (#304)

fal_v1.3.4

• fix: no default for deployment strategy (#297)
• feat(toolkit): better jpeg quality (#298)
• fix(deploy): tuple -> Tuple for 3.8 compatibility (#299)

fal_v1.3.3

• fix(sdk): default deployment strategy -> recreate (#296)

fal_v1.3.2

• feat: disconnect handler (#286)
• feat: fal file (#281)
• feat(fal-file): support run command (#290)
• feat: send strategy in app register (#293)
• fix(deploy): only use get_endpoints() for Apps (#294)

isolate_proto_v0.5.3

• feat: add deploy strategy to isolate proto (#292)

isolate_proto_v0.5.2

• proto: add machine_types and num_gpus option (#288)

fal_v1.3.1

• ci: add retry for failing test (#275)
• chore(container): Enable tests (#237)
• chore: bump timeout to health check request (#276)
• feat(toolkit): support multipart uploads (#278)
• feat(deploy): show available endpoints (#284)
• fix(field-exception): pydantic error message template (#282)
• fix(fal): don't forget to sys.exit when running as a module (#285)

fal_v1.3.0

• feat: add custom app exception handling & auto CUDA OOM exception handling (#274)